Quick answer: Google has shown a way to cut errors in one of quantum computing’s biggest bottlenecks, the preparation of so-called “magic states”, by about 40 times. Using its Willow quantum chip, the team produced magic states with a record 99.99% fidelity. The results were published in the journal Science on 15 October 2026.
That sounds technical, so here is the simple version, and why it matters.
First, why quantum computers make mistakes
Ordinary computers use bits that are 0 or 1. Quantum computers use qubits, which can hold a mix of both at once. That lets them tackle some problems, like simulating molecules for new medicines, far faster than any normal computer.
The catch is that qubits are extremely fragile. Heat, vibration or tiny electrical noise can scramble them. So quantum computers group many physical qubits into one more reliable “logical” qubit and constantly check for errors. This is called quantum error correction.
What are magic states?
Error-corrected qubits can easily do some operations, but not all the ones needed for a fully useful quantum computer. To do the missing operations, the computer has to prepare special ingredients called magic states.
- They are hard to make and get used up after one use.
- Some estimates say up to 95% of a future quantum computer’s capacity could go into producing them.
- So making them cheaper and cleaner is one of the biggest problems in the field.
What Google demonstrated
Google used a method called magic state cultivation, proposed in 2024 by its researchers Craig Gidney, Noah Shutty and Cody Jones. Think of it as a filter:
- Start with a rough, noisy magic state in a small region of the chip.
- Run a series of coordinated checks to catch flawed attempts.
- Throw away the bad ones and keep only the very pure states that pass.
The results on the Willow chip:
- About 40x fewer errors than magic states made without cultivation.
- 99.99% logical fidelity, which Google calls a record.
- The team also moved a purified state from one type of error-correcting code to another, showing it survives being passed around inside the hardware.
Why it matters
Google describes this as the first hardware demonstration of a practical fix for the magic state bottleneck. If magic states become cheaper, a useful quantum computer could need far fewer qubits than expected, which brings practical machines closer.
Areas that could benefit first include drug discovery, new materials and batteries, and certain optimisation problems.
Is my phone’s encryption or Bitcoin at risk?
Not today. Breaking modern encryption would need a much larger, fully error-corrected quantum computer than exists now. But each step like this shortens the timeline, which is why governments and companies are already moving to post-quantum cryptography, new encryption designed to resist quantum attacks.
The caveats
- This was shown on a small scale. It hasn’t been proven on the large machines needed for real-world tasks.
- Google itself says significant technical challenges remain, and gives no timeline for a useful quantum computer.
Frequently asked questions
What did Google announce about quantum computing?
Google showed that “magic state cultivation” on its Willow chip cuts errors in magic states by about 40 times, reaching 99.99% fidelity. It was published in Science on 15 October 2026.
What is a magic state in quantum computing?
A specially prepared quantum state needed to perform operations that error-corrected qubits can’t do on their own. They are hard to make and are used up after one use.
What is Google Willow?
Google’s quantum processor, used by Google Quantum AI for error-correction experiments.
When will quantum computers be useful?
There is no firm date. This result is a step forward, but larger machines and more breakthroughs are still needed.
Sources: Google Blog; The Quantum Insider; Science (15 October 2026).
